Research map / Optimization / Optimizers
Low-Variance Black-Box Gradient Estimates for the Plackett-Luce Distribution
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · ER1, 10 nodes | SHD: 2.2$\pm$2.9 -> 5.2$\pm$2.5 | Hurt |
| Use PL-RELAX rather than SINKHORN_ECP | Causal structure learning by order search · ER4, 10 nodes | SHD: 31.2$\pm$2.6 -> 29.4$\pm$1.9 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF1, 10 nodes | SHD: 0.8$\pm$0.4 -> 2.2$\pm$1.5 | Hurt |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF4, 10 nodes | SHD: 5.0$\pm$9.5 -> 8.2$\pm$3.1 | Hurt |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · ER1, 20 nodes | SHD: 18.6$\pm$13.5 -> 14.4$\pm$5.3 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · ER4, 20 nodes | SHD: 103.4$\pm$10.9 -> 71.0$\pm$5.9 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF1, 20 nodes | SHD: 2.0$\pm$1.4 -> 4.0$\pm$0.6 | Hurt |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF4, 20 nodes | SHD: 50.6$\pm$31.5 -> 20.0$\pm$4.3 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · ER1, 50 nodes | SHD: 38.2$\pm$21.6 -> 19.2$\pm$6.9 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · ER4, 50 nodes | SHD: 525.6$\pm$35.5 -> 220.6$\pm$42.8 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF1, 50 nodes | SHD: 38.8$\pm$39.3 -> 11.4$\pm$3.3 | Helped |
| Use PL-RELAX rather than GREEDY-SP | Causal structure learning by order search · SF4, 50 nodes | SHD: 381.2$\pm$76.2 -> 70.0$\pm$9.9 | Helped |
Closest papers by meaning
- REBAR: Low-variance, unbiased gradient estimates for discrete latent variable models (NeurIPS 2017 · similarity 0.87)
- Using Large Ensembles of Control Variates for Variational Inference (NeurIPS 2018 · similarity 0.86)
- Efficient preconditioned stochastic gradient descent for estimation in latent variable models (ICML 2023 · similarity 0.86)
- A Stochastic Composite Gradient Method with Incremental Variance Reduction (NeurIPS 2019 · similarity 0.85)
- Marginalized Stochastic Natural Gradients for Black-Box Variational Inference (ICML 2021 · similarity 0.85)
- Gradient Estimation with Discrete Stein Operators (NeurIPS 2022 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.